Abdominal aortic aneurysms (AAA) are pathological dilations in local aortic wall. The inflammatory infiltrates of the perivascular adipose tissue (PAT) surrounding AAAs were associated with AAAs and have been shown to contribute vascular pathology. However, the mechanism by which PAT inflammation contributes to vascular pathology in AAA remains to be clarified. This study aimed to explore the association between immune cell infiltration and key gene expression profile in PAT of AAA. For that, a gene expression dataset of human dilated perivascular adipose tissue (dPAT), non-dilated perivascular adipose tissue (ndPAT), subcutaneous abdominal fat (SAF) and omental-visceral fat (OVF) samples, as well as another microarray dataset of the abdominal perivascular adipose tissue in peripheral artery disease patients were downloaded from GEO database for analysis in this study. The CIBERSORT algorithm, weighted gene co-expression network analysis (WGCNA) and LASSO algorithm were used for the identification of immune infiltration, immune-related genes and the development of diagnostic signature. Our data discovered a significant higher proportion of activated mast cells and follicular helper T (Tfh) cells in dPAT than ndPAT, OVT and SAF samples. Moreover, AP-1 family members (FOS, FOSB, ATF3, JUN and JUNB) were found to compose the hub genes of purple module in WGCNA. Among them, FOS gene acts as a higher efficient marker to discriminate dPAT from ndPAT, OVT and SAF in AAA. Meanwhile, the expression profiles of the AP-1 family members are all significantly positive correlated with activated mast cell, plasma cell and Tfh cell infiltration in dPAT of AAA. Therefore, in the PAT surrounding AAA, the signature of inflammatory infiltration might be represented by a FOS-dominated cell network consist of activated mast cell, plasma cell and Tfh cell. Given the complicated etiology of AAA, our results are likely to shed new light on the pathophysiologic mechanism of AAA influenced by the local dPAT.